Knowledge-Enhanced Causal Discovery Methods
Time: 2026-09-17
Published By: He Liu
Speaker(s): Danlei Gu (BICMR)
Time: 15:00-16:00 September 18, 2026
Venue: Room 77201, Jingchunyuan 78, BICMR
Causal discovery aims to identify causal relationships between variables from observational data; however, traditional data-driven methods often underutilize prior domain knowledge, facing challenges in both identifying causal relationships and interpreting underlying mechanisms within complex systems. This presentation introduces a knowledge-enhanced causal discovery method that integrates knowledge graphs and large language models into the discovery process. This approach generates candidate causal relationships and mechanistic hypotheses that are consistent with domain knowledge, subsequently employing data-driven methods for causal verification. By synergizing domain knowledge with observational data, this method offers a novel approach to uncovering causal mechanisms in complex systems.
